Cell Charge-Discharge Efficiency testing

LEDSchlucker

Member
Joined
Dec 20, 2018
Messages
157
Hi guys,

a while ago I asked myself what the charge-discharge efficiency of a li-ion cell is. So I tried to measure it with an Opus, but the result that came out is pretty disappointing...


Measured Values:

Cell 1: Cell 2: Cell 3: Cell 4:
CH 2115mAh 2106mAh 2035mAh 2086mAh
DC 2378mAh 2105mAh 2097mAh 2130mAh

CH 2151mAh 2045mAh 2083mAh 2060mAh
DC 2394mAh 2071mAh 2054mAh 2087mAh
CH 2219mAh 2018mAh 2061mAh 2019mAh

DC 2451mAh 2125mAh 2117mAh 2134mAh
------------------------------------
Average per cell:
CH 2162mAh 2056mAh 2060mAh 2055mAh
DC 2408mAh 2100mAh 2089mAh 2117mAh
------------------------------------
------------------------------------

Total average:
CH 2083mAh
DC 2179mAh

"efficiency" (measured with Opus) = 2179mAh / 2083mAh = 1.046 which is 104.6% (which is impossible, because you can't take more electricity out than you put in)


As the result is over 100% (104.6%), the Opus has to be very imprecise! :s



Recently I read somewhere that the Opus is adding 30% to the discharged value because of the high impedance of the springs.
Is that correct, does the Opus add 30% to the measured capacity?
 
Not 30 but between 5 and 10 is what it is off. Especially att 1a discharge. Wolf Cherry and crew here have some threads showing it.
 
LEDSchlucker said:
Recently I read somewhere that the Opus is adding 30% to the discharged value because of the high impedance of the springs.
Is that correct, does the Opus add 30% to the measured capacity?

The Opus is not that bad on the C/D/C cycle (I testat 500mA) The discharge mAh rating is slightly elevated as Daromer said. Max I would say is 5% inflation. The IR readings is where the discrepancy comes in and that is where the number 30 stuck in your head. You need to take a 10% to 20% inaccuracyoff of IR and a 30m? deduction from the reading into account. It is in the owner's manual such as it is. Page 3 on the bottom under Quick Test.

Wolf
 
Ahh, that was IR...
But still strange that the Opus measures DC capacity higher than CH capacity. That makes somehow no sense to me.

Just did a quick and dirty Fan Mod and on one slot I did a wire Mod. Let's see what the difference in cap will be.
 
Because its a cheap tester. Its a simple algoritm where it interpolates the values and not callibrated. Use the number as number and only on DC. So not use it as real capacity ;) atleast not with extensive testing and knowledge
 
I know that this is the simple reason, but if you think deeper, the software has to change something on the values, otherwise it wouldn't be possible to get higher reading on DC than on CH.

Because while charging the Opus measures higher voltage than on the battery and when discharging it measures lower than real, because of the resistance in the springs etc. This would cause a drift in the other direction!
 
I suspect folks are expecting a little to much from such a low cost item, they are fine for comparative mass testing, but not great for overly accurate or repeatable results.

I view those testers in the same way as I view neon screwdrivers.

If someone such as Wolf wants to do some analysis on the affect contact resistance/drop testing under load has using any of the common testers in combination with a better quality, more repeatable set of contacts such as the type depicted below I'll happy buy the tester and external contacts for them.

It would be interesting to see the tolerance of the resistorsused during the discharge phase - replacing that with a high tolerance item might also increase the accuracy.


image_jybckp.jpg
 
I know for precise measurement I have to take a four wire tester.

What type of discharge tester do you recommend for precise measurements?
Or does anyone know a really precise ADC, so I could build my own tester?
 
LEDSchlucker said:
I know for precise measurement I have to take a four wire tester.

What type of discharge tester do you recommend for precise measurements?
Or does anyone know a really precise ADC, so I could build my own tester?

I have a large, expensive Cadex, but that's for my commercial work, probably not what a hobby or DIYer would need or want to pay for.

I suspect the software and control algorithms used within the sub 50 testers are capable of very good accuracy, but the results can only be as accurate as the components used, unless thereis a process to calibrate them (and the calibration process is only as accurate as the equipment used to calibrate against)

Just as a quick test, perhaps someone with a known accurate ammeter couldtest the accuracy of the discharge current of their favourite charger/discharge tester - perhaps do a discharge test the old fashioned way by timing the discharge.

Is the discharge current constant across the full voltage range of the test ? (I suspect not)
 
Sean said:
I suspect folks are expecting a little to much from such a low cost item, they are fine for comparative mass testing, but not great for overly accurate or repeatable results.

I view those testers in the same way as I view neon screwdrivers.

Hey hey I have neon screwdrivers. Snap on too."Expensive". :p

image_egbgvx.jpg


If someone such as Wolf wants to do some analysis on the affect contact resistance/drop testing under load has using any of the common testers in combination with a better quality, more repeatable set of contacts such as the type depicted below I'll happy buy the tester and external contacts for them.

It would be interesting to see the tolerance of the resistorsused during the discharge phase - replacing that with a high tolerance item might also increase the accuracy.


image_jybckp.jpg
Sean
Hmm,
I am intrigued Ialready have a set of 4 good contact cell holders

image_chtpvd.jpg

I was going to use them on my ZB206 testers but too cold in the shed to build right now. They are single units but similar as to the 4 wire ones you are showing.
So that would be a good start. I would assume the one to start with is an Opus.Replace the sliders with the holders depicted above properly soldered to the Opus PCB. (I wonder if they have a separate circuit tap to measure IR?) The load resistors would be pretty easy to changejust got to find the value.
That would be a Super Mod on the Opus. It would be more of a fun project as financially feasible it is not. More like putting a V10 motor on a Go Cart.
Might as well buy a SKYRC MC3000.
Sure I'm game. Let's do it.
Let me know

Wolf
 
The MC3000 looks to be just another clone of ? - they all look the same to me, likely running very similar firmware.

If everyone can agree on the best or most populartester (Opus ?) and a set of those 4 way/4 wire sets of contacts that are readily available from a trusted source, I'll fund them so Wolf can experiment with the aim of improving the accuracy and repeatability of "cost effective" testers.
 
LEDSchlucker said:
I know for precise measurement I have to take a four wire tester.

Yep
YR1030 very reasonably priced and very consistent.
Review of it herehttp://budgetlightforum.com/node/56582
His reference was theKeithley DMM7510 to see how good it was.

What type of discharge tester do you recommend for precise measurements?
I just recently got a SKYRC MC3000 and I like it a lot. I am presently doing an eleven cycle CDC on 4 sony "green" suck cells for thethread.
Nice unit very programmable heavy and tough.

For Sean
Not cheep $110.00
And believe you me no clone of the firmware of the cheap ones also has the ability to reflash with new firmware if it gets updated.
Graphing PC interface etc all good stuff.


Wolf
 
Isnt this a MC3000 ? - looks very similar to every other cheap tester.

I'll have a read of that link later.


image_umdfxh.jpg
 
YR1030
Wolf said:
Yep
YR1030 very reasonably priced and very consistent.
Review of it here http://budgetlightforum.com/node/56582
His reference was the Keithley DMM7510 to see how good it was.
I'll probably have to buy one...

Is the ZB206 precise? I know it got extra a voltage measurement connector, but just to be sure, can I trust it?


Sean said:
Isnt this a MC3000 ? - looks very similar to every other cheap tester.

I'll have a read of that link later.


image_umdfxh.jpg

The MC3000 doesn't look cheap to me. Tell me which cheap (<$50) Tester is able to connect to the PC and see graphs?
 
LEDSchlucker said:
YR1030

I'll probably have to buy one...
Is the ZB206 precise? I know it got extra a voltage measurement connector, but just to be sure, can I trust it?
I personallydon't knowif the ZB206 is accurateas I have not built my testing board with them yet. There are some reviews of it out there but I always like to verify myself.It has been put on the back burner as for 2 reasons I have the YR1030 which is a great testerand it's too cold to work in the shed.As soon as I get one up and running though I will report the results.

Wolf


Sean said:
Isnt this a MC3000 ? - looks very similar to every other cheap tester.
I'll have a read of that link later.
Here is the link to the infamous test of the SKYRC MC3000

https://lygte-info.dk/review/Review Charger SkyRC MC3000 UK.html
From the review
"Conclusion
This charger can handle many chemistries and many battery sizes and in addition to this it has a good precision, all of this is way above any other (consumer) charger on the market at the current time."

Wolf
 
That MC3000 certainly does appear to have a lot more features than the 50 testers, but the cell contacts dont look much of an improvement over the cheaper ones and I wonder how accurate and repeatable it really is.

Perhaps, given its ability to data record that's what we should be trying to improve - with multi point contacts, higher tolerance discharge resistors, perhaps there areaccurate current sensors already built in etc etc

Is trying to make a silk purse out of a sows ear rather pointless ?remembering that most of you are wanting a quick go/no go method of elimination, along with a +/- 10% test of capacity so that cell packs can be formed.
 
Guys, i have to call you to order:
Opus et al. are CHARGERS.
They are no measurement equipment. They have not at all be designed for ANY repeatability and/or accuracy.
Examples: The sliders, the springs, "normal" (5 %) resistors ( no accurate ones..).

Even the PROCESS of charging and discharging of a cell does not give accurate results, see:

Let us assume a 1000mAh cell is discharged at 500 mA. that gives 7200 seconds. roughly 1 mAh every 12 seconds.

The voltage difference is 4.2 - 2.7 = 1.5 Volts. That gives roughly 50 seconds per mV decrease or increase. a 1 mV "Error" gives about 5 mAh difference.
( i know the slope at both ends of the curve is more steep, and the values would change by , say, about faktor 5 ).

Now, Voltage measurements fights for accuracies of 5 mV or more, especially under the circumstances that voltage converters are freely running on the board.
 
Cherry67 said:
Guys, i have to call you to order:
Opus et al. are CHARGERS.
They are no measurement equipment.

Of course they are "measurement equipment" - they run discharge tests and produce a capacity result (of questionable accuracy)

The possible project aim would be to increase the accuracy of the result, in a cost effective manner.

A charger, just charges.
 
Sean said:
That MC3000 certainly does appear to have a lot more features than the 50 testers, but the cell contacts don't look much of an improvement over the cheaper ones and I wonder how accurate and repeatable it really is.

Sean,

Having laid my hands on one and actually owning one I must say that the contacts are extremely robust and it takes a strong man and a small boy just to get a cell in there. The sliders are very and I mean very thick. I will get my vernier calipers out and check the thickness. It does notmake it a better analyzer by any stretch of the imagination but we will see. Maybe they just clamored some heavy metal together wrote some reasonable firmware and called it goodThat being said I have not opened it up yet to see what's inside.I can tell you the IR readings are not that goodbut then again only a 4 wire will give you the true IR. The repeatability is in the works right now as its running an eleven cycle test on some green sonys. When that's done I will run a cycle test on my new INR18650-25R and see.

Wolf
 
For me works an Opus to sort cells and eliminate the bad ones.

What I'm searching now is an accurate measuring device, which is precise enough to compare charged energy to discharged energy to get the efficiency of a cycle.

Maybe I have to go bigger and use 10 or even 20 cells in parallel for that.
 
Back
Top